Capacitor testing

[stevez]

My friends and I tend to make things from our junkbox collections of parts. The condition of these parts is questionable at times and it would be nice to test, to the extent possible, for some easy things. I can test some of the capacitors with my DVM but that only tells part of the story, I think. When I had a VOM I could test electrolytics with the ohm-meter scale and watch the capacitor charge.

It would seem that applying a DC voltage might tell some of the story. A friend thought of using a curve tracer though those are hard to come by, especially at higher voltages.

Any thoughts or ideas would be appreciated.

 

[Johnson777717]

Have you thought about using a capacitance meter? Maybe an ESR meter for electrolytics?

I'm currently re-designing a method for a capacitance meter, which will be posted at:
www.electronicprojects.org

The meter measures caps ranging from 2pF up to 2uF, and displays the capacitance on a 4 digit 7 seg display.

Some place sells ESR meters (for electrolytics) or a kit, I'm sure if you search for ESR meter, you'll find a bunch of items.

This site uses an AVR for measuring capacitance, if you're up to speed on AVR stuff.
http://elm-chan.org/works/cmc/report.html

 

[stevez]

I probably wasn't as clear as I could have been in my first post. I am not sure that by measuring the value of the capacitor alone, that it will tell me the whole story about the capacitor. Certainly if it has shorted the test will (or should) reveal that. I was thinking more of the catastophic failures that I and others have experienced. The capacitor appeared to be ok but blew up when stressed at the normal operating voltage of the circuit.

A friend, with more knowledge of this stuff recalls using a curve tracer to see at what point breakdowns begin to occur.

I had wondered if applying the rated DC voltage across the terminals while monitoring what I'll call leakage current wouldn't tell part of the story. I'd use a very high value resistor to limit current to a trickle which might require a brief wait for the capacitor to charge. Not sure I'd know how to recognize a good one or bad one just yet. My thought process applies to the kinds of stuff that I use - generally 50 working volts DC or less - on occasion 100 volts.

 

[Nigel Goodwin]

I probably wasn't as clear as I could have been in my first post. I am not sure that by measuring the value of the capacitor alone, that it will tell me the whole story about the capacitor. Certainly if it has shorted the test will (or should) reveal that. I was thinking more of the catastophic failures that I and others have experienced. The capacitor appeared to be ok but blew up when stressed at the normal operating voltage of the circuit.

What sort of capacitors?, Tantalums are usually the ones which blow up, electrolytics do as well - but not as often, nor as violently. Certainly if you exceed their voltage rating they don't like it!.

I had wondered if applying the rated DC voltage across the terminals while monitoring what I'll call leakage current wouldn't tell part of the story. I'd use a very high value resistor to limit current to a trickle which might require a brief wait for the capacitor to charge. Not sure I'd know how to recognize a good one or bad one just yet. My thought process applies to the kinds of stuff that I use - generally 50 working volts DC or less - on occasion 100 volts.

I've got a very old (very, very old!) item of test equipment at work, made by Pye - it includes RF and AF generators, watt meter, and a capacitance bridge - including a leakage tester. I've only ever used it for testing and measuring capacitors, the leakage section allows you to switch various voltages across the capacitor under test, up to a maximum of 440V.

The suggestion of an ESR meter is a good one, although it won't catch exploding capacitors it finds most faulty electrolytics - I built one a year or two back, it's been invaluble. One of the biggest reliability problems over recent years is low-ESR electrolytics.

 

[Optikon]

I probably wasn't as clear as I could have been in my first post. I am not sure that by measuring the value of the capacitor alone, that it will tell me the whole story about the capacitor. Certainly if it has shorted the test will (or should) reveal that. I was thinking more of the catastophic failures that I and others have experienced. The capacitor appeared to be ok but blew up when stressed at the normal operating voltage of the circuit.

A friend, with more knowledge of this stuff recalls using a curve tracer to see at what point breakdowns begin to occur.

I had wondered if applying the rated DC voltage across the terminals while monitoring what I'll call leakage current wouldn't tell part of the story. I'd use a very high value resistor to limit current to a trickle which might require a brief wait for the capacitor to charge. Not sure I'd know how to recognize a good one or bad one just yet. My thought process applies to the kinds of stuff that I use - generally 50 working volts DC or less - on occasion 100 volts.

Pulling parts out of a junk box, two concerns come to mind:

1) The part was in use and possibly overstressed (it was removed from a circuit as a suspect part and later found to be ok and thrown in the junk box) but it could have been over stressed (this is key because it could be at the end of its useful life)

2) The part was never overstressed and perhaps not even used at all but it may have been in there for a very long time.

One would like to test these with some confidence that they are still ok (or not) for general purpose use.

In the realm of electrolytics, I believe that in the long term, the electrolyte evaporates or somehow chemically degenerates and the properties of that capacitor will change. One cannot rely on only the capacitance value (as it may have had a wide tolerance to begin with)
Although, if one knows the tolerance, and the part is found to be outside, why guess about it? I would just throw it out (could save lots of time later on during testing when you've long forgotten about that "shady" part you used.)

Some of the properties I would expect to change would be:

1) leakage current
2) Dielectric absorption
3) Breakdown voltage
4) ESR
5) Nominal capacitance

A voltage breakdown test could also be part of a D.A. test, because you have applied a known voltage for a known time and this is exactly the test used to characterize D.A.

Leakage current could be used with a DC voltage.

ESR could be done by applying a step input with the capacitor in series with a known inductance. From the overshoot / ringing phenomenon, one can ascertain the ESR and the capacitance as well. A current pulse could be used as well. You would have to provide a scope hookup however to see the time domain reaction.

For quick and dirty, some of these tests are impractical I think. I also think with an unkown capacitor under test, the best you will be able to do is compare the result with a nominal one for that type. You may want to consider having some charts handy that show what a nominal parameter might look like for a given type so you can make a very quick assessment and move on.

thats my .02

 

[Russlk]

Electrolytic caps loose voltage rating over time if not in use, but can be re-formed by passing a current thru them. Connect the electrolytic to a supply of rated voltage, in series with a current limiting resistor which limits the current to a few milliamps. When the current reduces to the rated leakage, it is ready to use.

 

[Dean Huster]

RC Bridge

In the "olden" days (1950-1970) of electronics, most kit makers had RC Testers or RC Bridges, instruments that measured resistance and capacitance and capacitor leakage at the rated voltage. Typical of these was the Knight-Kit KG-670 and the EICO 950B. They had a "magic eye" tube as the indicator for bridge balance (when measuring capacitance) and for the leakage test. Most of these won't go below about 50v, however since they were designed during the heyday of vacuum tubes. Still, for caps of 50v rating and up, they're great for most tests. You can also use them to reform capacitors. The original kits usually sold for around $25, indicating that you can get them for a song on ebay. Don't expect them to be as accurate as a digital cap meter or as sensitive in seeing leakage as a semiconductor curve tracer. In fact, any that you'll buy will probably have to have its own caps replaced, because they'll be old paper types that are leaky as a helium balloon. But if you replace them, especially the bridge caps, with close tolerance Mylar caps, you'll have an instrument that is more accurate than when it was made.

Dean

 

[Nigel Goodwin]

ESR could be done by applying a step input with the capacitor in series with a known inductance. From the overshoot / ringing phenomenon, one can ascertain the ESR and the capacitance as well.

ESR is generally specified (and therefore requires measuring at) 100KHz, the ESR meter I built uses a 100KHz sinewave oscillator along with a AC milli-volt meter to measure it's impedance at 100KHz. It uses a very low level of signal, so can happily test them in circuit.

 

[stevez]

Thanks! Excellent information and ideas. I am sure that others can contribute more but it sounds like I could do three things to increase the likelihood that I have a good capacitor.

Measure the value and compare to the value marked.
Measure ESR - I'd have to learn how to interpret if I don't have nominal value.
Apply rated voltage at low current - to see if it holds up and possibly to "reform"

While these tests won't necessarily prove that a capacitor will hold up it does increase the likelihood that I'll have a good part.

Implied in Optikon's comments and worth reiterating - there's a degree of uncertainty with used parts or junk box parts. For beginners I've made the suggestion that they purchase a few new parts rather than bags full of discards or junk. It can add a lot of confusion to an already confusing situation - when you are a beginner or even when you've been around a while. I've found that if what I am building is at all tricky or unfamiliar I try to stick with parts of known condition. Still, when I am fooling around and experimenting that junk box can be a good source of low cost parts.

Again, great stuff - thanks.